I just wrapped up what was honestly the most concentrated stretch of conversations with network engineers I’ve had all year.

NetBrain LIVE brought together customers, partners, and prospects across a full range of environments and team sizes. And while there was plenty of great technical content on stage, the conversations that stuck with me happened in the hallways. At the tables. In the side rooms where people say what they actually think.

One topic kept coming up no matter who I was talking to.

AI. And whether to trust it.

Three Types of Engineers in Every Room

I didn’t plan for this to be a theme. But by the third or fourth conversation it was impossible to ignore.

Every team I talked to fell somewhere on the same spectrum.

Some engineers were already using AI in their workflows. Not in a theoretical way, in a “I ran this through it before I got on the bridge call” way. They were using it for context. For pattern matching. For getting a faster read on a problem they’d seen variations of before. And they were quietly impressed with what it could do.

Some engineers wanted to use it but couldn’t. Not because they didn’t see the value…they did, but because their organization hadn’t figured out the policy yet. Security concerns. Data sovereignty questions. Leadership that wasn’t ready to sign off. The capability was sitting right there and they were being asked to wait.

And some engineers were skeptical. Not hostile, just unconvinced. They’d seen enough vendor promises not to take the pitch at face value. They wanted proof that this wasn’t just another layer of tooling that would generate more noise without actually making operations easier.

What struck me wasn’t that these three groups existed. It was that they were all in the same industry, often at organizations of similar size and complexity, and they were in completely different places on this.

The Tools Are Here. The Trust Is Still Being Built.

Here’s the thing I kept coming back to in those conversations.

We’ve been through tool cycles before. Monitoring got smarter. Automation frameworks arrived. Dashboards multiplied. Each wave came with a version of the same promise: less manual work, faster resolution, better visibility.

And teams invested. They implemented. Some of it worked. Some of it created new overhead without solving the underlying problem.

So when AI shows up with a similar pitch, the skepticism makes sense. It’s earned.

But what I heard from the engineers who were actually using it was something different from previous tool cycles. They weren’t talking about dashboards or alert volumes. They were talking about context. About getting a faster answer to “have we seen this before and what did we do about it.” About not having to track down the one person on the team who remembers how that particular environment behaves.

That last part is what caught my attention.

Because the problem they were describing wasn’t really an AI problem. It was a knowledge problem. The operational intelligence that drives resolution, the patterns, the institutional memory, the understanding of how this specific network behaves under pressure. That stuff still lives primarily in people’s heads. Specific people. And AI was starting to feel, to some of them, like a way to finally do something about that.

The Tension That’s Been There All Along

I’ve been in and around network operations long enough to know this tension isn’t new.

Teams have more tooling than ever and operations still feel harder than they should. Alerts fire. Engineers mobilize. The same handful of people end up on the bridge call because they’re the ones who actually know what to do. And when something breaks in a way that’s slightly different from last time, you feel the gap between what your monitoring tells you and what you actually need to know to fix it.

• Monitoring is not the same as understanding.
• Automation is not the same as operational change.
• And visibility, it turns out, is not the same as control.

What I think I was watching at NetBrain LIVE was the beginning of a real conversation about that gap. Not the vendor version of the conversation. The practitioner version. Engineers talking honestly about where the limits are and whether AI might finally be the thing that helps close them, not by replacing judgment, but by making institutional knowledge something more than tribal.

Where This Is Going

I don’t have this fully mapped out yet.

But I’ve been thinking a lot about the journey network operations teams actually go through. The stages. The moments where progress stalls. What it takes to move forward when the tooling is good but the operation is still stuck.

I’ll be writing more about that in 2026. Some of it is still forming. But the conversations at NetBrain LIVE made me more confident that the timing is right. Practitioners are ready to have a more honest conversation about where network operations is actually headed and what it’s going to take to get there.

If your team is somewhere on that AI trust spectrum right now; fully in, waiting on policy, or still skeptical…I’d genuinely like to hear where you are. Because what I heard this year tells me the experience is pretty universal, even if the conclusions aren’t the same yet.

That feels like the right place to start.